AUTOMATIC ASSESSMENT OF TIME-RESOLVED OCT IMAGES FOR SELECTIVE RETINA THERAPY

摘要

The invention relates to a method for predicting a pattern in an optical coherence tomography sample image (OCT) data set (X), corresponding to a structural change in a retina (51) of an eye (5), comprising the steps of performing an OCT measurement of the retina (51), automatically providing a sample image data set (X) comprising a plurality of image data points (xi), automatically extracting a first plurality of features (u) from the sample image data set (X), and automatically predicting the presence or absence of the pattern from the first plurality of features (u) by means of a classification algorithm, wherein a second plurality of features (u) extracted from a training image data set serves as a training data set of the classification algorithm. The invention further relates to a system, and a computer program for predicting a pattern in an optical coherence tomography sample image data set (X).